集成预报在巴彦淖尔市气温预报中的应用  

The Application of Consensus Forecast in Temperature Forecast in Bayan Nur city

作　　者：张玉荣[1] 王永清[1] 高玲[1] 梁凤娟[1] Zhang Yurong WangYongqing Gaoling LiangFengjuan(Bayan Nur Meteorological Bureau, Linhe, 015000)

机构地区：[1]巴彦淖尔市气象局,内蒙古临河015000

出　　处：《山西气象》2016年第3期7-10,共4页Shanxi Meteorological Quarterly

摘　　要：本文针对欧洲模式（Ec）、中国模式（T639）、德国天气在线、美国天气、中国天气、中央气象台指导预报六个模式预报中的日最高气温和最低气温预报值，使用LINEST函数做多元回归分析，得出六家模式的集成预报结果，结果表明：集成预报比六家模式的预报准确率都高；同时，建立本地化订正方法，将原六家模式最高最低气温预报值进行订正后再集成，结果表明：订正后的集成预报比直接集成预报准确率又提高了0～4％，说明订正方法的使用对提高最高、最低气温预报准确率有一定的效果；另外，通过对各家模式预报结果和集成预报结果的检验分析，不仅为预报员择优使用数值预报产品提供参考依据，也为研究数值预报产品释用提供一定的参考方法。Based on the maximum and minimum temperature forecast by EC model,T639, WetterOnline, AccuWeather, Weather China and National Meteorological Center of CMA, the consensus forecast results of the six models was got using the multiple linearregression method based on the LINEST function. The resuhs showed that：the accuracy of the consensus forecast was higher than the six models; if the maximum and minimum temperature of the six models corrected by the local correction method, the accuracy rate of the consensus forecast increased by 0 - 4%, which indicated that correction method could improve the forecast accuracy of maximum and minimum temperature; the analysis of the verification provided forecasters reference of the interpretation of numerical weather forecasting products.

关 键 词：最高最低气温预报 LINEST函数 数值预报产品 多元线性回归 

分 类 号：P456.9[天文地球—大气科学及气象学]